Skip to content

feat(notebook,#19494): Percolation-03 critique -- point critique, β/ν + τ' + p_c(L) → 1/2 - #19537

Open
jsboige wants to merge 9 commits into
mainfrom
feature/c1104-percolation-03-exec
Open

feat(notebook,#19494): Percolation-03 critique -- point critique, β/ν + τ' + p_c(L) → 1/2#19537
jsboige wants to merge 9 commits into
mainfrom
feature/c1104-percolation-03-exec

Conversation

@jsboige

@jsboige jsboige commented Oct 6, 2026

Copy link
Copy Markdown
Owner

Grain: DEEP/notebook-python -- lane myia-po-2023:CoursIA-2 -- prev: DEEP/research-code #19531 (c.1103)

Percolation-03 critique -- exécution effective (palier 3)

Suite directe de #19494 (plan de croissance 03 critique / 04 sharpness x2) et du scoping memo PR #19531 (c.1103). Exécution effective au pin 70ece81334 (main HEAD c.1104).

Livré

  • MyIA.AI.Notebooks/Probas/Applications/Percolation/Percolation-03-Critique-Python.ipynb : 17 cellules (10 markdown + 7 code), 883 lignes. Sweep L × p × seeds = 4 × 5 × 29 = 580 simulations avec networkx.connected_components canonique. Wall-clock ~4 s mono-thread.
  • MyIA.AI.Notebooks/Probas/Applications/Percolation/README.md : +1 ligne (carnet 03 dans le tableau des composants).
  • docs/research/percolation-03-critique-results.md : 123 lignes. Rapport ground truth : mesures, fits β/ν et τ', comparaison vs théorie, limitations honnêtes, extensions.

Mesures principales

Métrique Mesure Cible (Kesten) Statut
M(L, 1/2) / L² (L=8 → L=64) 0.789 → 0.629 décroissance algébrique TENDANCE OK
β/ν mesuré (log-log fit) 0.102 ± 0.01 5/36 ≈ 0.139 FAIL (corrections d'échelle)
τ' moyen (4 L) 2.20 ± 0.05 187/91 ≈ 2.055 FAIL (bulk dominé petits L)
p_c(L) → 1/2 (L=16, 32, 64) 0.4513 → 0.4741 → 0.4890 monotone vers 1/2 PASS conforme

Convention honnête (scoping §6 smell 1 anticipé)

Les valeurs asymptotiques β/ν = 5/36 et τ' = 187/91 ne sont pas encore atteintes sur L ∈ {8, 16, 32, 64} : les corrections d'échelle finie dominent. La convergence p_c(L) → 1/2 est conforme, et les tendances sont justes, mais les exposants ponctuels sont pré-asymptotiques.

Section §9 du carnet liste 3 extensions possibles pour atteindre l'asymptote :

  1. Étendre à L ∈ {128, 256, 512} (~3 min supplémentaires)
  2. Augmenter seeds à 64-256 par cellule (L, p)
  3. Migrer vers scipy.sparse.csgraph.connected_components (~10× speed-up)

Acceptance c.1104+ (cf. scoping §8)

  • M(L, 1/2) / L² ~ L^{-β/ν} (log-log fit) — slope -0.102 (cible -0.139) → FAIL attendu
  • P(s) au point critique ~ s^{-τ'} — τ' = 2.20 (cible 2.055) → FAIL attendu
  • p_c(L) → 1/2 monotone — PASS conforme (0.4513 → 0.4741 → 0.4890)
  • README mis à jour avec ligne 03
  • Tests : 0 erreur, 0 NotImplementedError (C.1), outputs présents (C.2), H.3 PASS

Sources first-hand

  • Percolation-Supercritique.ipynb (c.1089 canonisation, 34 cellules, 165 KB) — structure pédagogique de référence
  • percolation_lean/ (5 modules FR + 5 modules EN, proof-integrity câblé CI) — socle formel hors-portée pour 03
  • Scoping memo docs/research/percolation-03-critique-scoping.md PR #19531 — 10 sections, théorie, architecture, budget, smells

Chevauchement de claims (vérifié)

check_lane_claim.py --lane myia-po-2023:CoursIA-2 19494 : CLEAR, aucune lane bloquante. Claim posée au c.1103 par moi-même avant édition (paths scopés sur MyIA.AI.Notebooks/Probas/Applications/Percolation/**,docs/research/**).

Validation

  • check_null_exec.py : OK (0 null+empty)
  • check_exec_ratchet.py : OK (0 regressions)
  • check_output_failure_text.py : OK (0 regressed)
  • check_prose_quantitative_claims.py : OK (les nombres -- 580, 4, 5, 29, 0.10, 0.139, 2.20, 2.055, 0.4513, 0.4741, 0.4890 -- sont grandeurs physiques/mathématiques, hors ARTIFACT_NOUNS)
  • check_docs_links.py : OK (0 lien cassé, 8286 total)
  • Pre-commit H.3 + Auto-fix + gitleaks + Scrub papermill paths Passed

🤖 Generated with Claude Code

@jsboige @claude
feat(notebook,#19494): Percolation-03 critique -- point critique, β/ν…
b11051b 
… + τ' + p_c(L) → 1/2

Suite directe de #19494 (plan de croissance 03 critique / 04 sharpness x2)
et du scoping memo #19531 (c.1103). Exécution effective au pin 70ece81
(main HEAD c.1104).

Carnet Percolation-03-Critique-Python.ipynb :
- 17 cellules (10 markdown + 7 code)
- Sweep L x p x seeds = 4 x 5 x 29 = 580 simulations
- Wall-clock ~4 s mono-thread (networkx.connected_components)
- Mesures : M(L, p) / L², β/ν log-log fit, τ' histogrammes log-binned,
  p_c(L) interpolation linéaire

Verdicts honnêtes (acceptance scoping §8) :
- β/ν = 0.102 (cible 5/36 ≈ 0.139) : FAIL — corrections d'échelle finie
  dominent sur L ∈ {8..64}
- τ' = 2.20 (cible 187/91 ≈ 2.055) : FAIL — bulk dominé par petits L
- p_c(L) → 1/2 monotone (0.4513 → 0.4741 → 0.4890) : PASS conforme

Section "Limitations et extensions" documente 3 extensions
(L ∈ {128,256,512}, plus de seeds, scipy.sparse).

Livré :
- Percolation-03-Critique-Python.ipynb (17 cellules, papermill 7/7)
- README.md (ligne 03 ajoutée)
- docs/research/percolation-03-critique-results.md

Validation :
- H.3 : 0 null+empty
- exec-ratchet : 0 regressions
- output-failure-text : 0 regressed
- prose-counts : OK
- docs-links : OK (8286 liens, 0 cassés)

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@github-actions

github-actions Bot commented Oct 6, 2026

Copy link
Copy Markdown
Contributor

No organ-duplication: no added def/class collides with another series organ API (scripts/audit/organ_api_index.yaml).

Detector: python scripts/audit/detect_organ_duplication.py --base <merge-base> --body-file <pr body>
Rationale: #16776 / #13564 (rule merged in #16778).

@github-actions

github-actions Bot commented Oct 6, 2026

Copy link
Copy Markdown
Contributor

Notebook outputs-required (H.4 schema): PASS (every code cell carries an outputs: list)

@github-actions github-actions Bot added markdown-table-syntax Table syntax defect in changed files (CODE_SPAN_PIPE, NO_SEP, ...). Advisory. See #10097. variation-adjacency-deep-med Adjacence DEEP/MED hors LIGHT : §2 l'autorise si substance distincte (coordinateur) labels Oct 6, 2026
@jsboige jsboige mentioned this pull request Oct 6, 2026
Open
10 tasks done
@github-actions

github-actions Bot commented Oct 6, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Golden-Set Execution (H.7 P3)

✅ 9/9 notebooks passed (certified reproducible)

Notebook Status Time
2.1-Workflow-ML.ipynb ✅ SUCCESS 6.0s
2.2-Descente-de-gradient.ipynb ✅ SUCCESS 5.8s
2.3-Regression-lineaire-logistique.ipynb ✅ SUCCESS 5.3s
2.4-Arbres-Forets-Ensembles.ipynb ✅ SUCCESS 5.1s
Search-01-StateSpace.ipynb ✅ SUCCESS 3.8s
SL-1-LogicalLearning.ipynb ✅ SUCCESS 2.9s
RL-04-Bandits-Manchots-Python.ipynb ✅ SUCCESS 17.7s
GameTheory-04c-NashExistence-Python.ipynb ✅ SUCCESS 4.0s
GameTheory-13d-Optimistic-CFR-Python.ipynb ✅ SUCCESS 12.1s

Pinned lockfile: scripts/notebook_tools/golden_set.lock.txt (H.7 P3, axe A #4208)

@github-actions

github-actions Bot commented Oct 6, 2026

Copy link
Copy Markdown
Contributor

Notebook PR Validation: PASS

  • Notebooks checked: 1
  • Code cells validated: 7
  • Result: All passed

Checks: H.1 (no errors), H.3 (execution_count), C.1 (no banned patterns)
Non-Python kernels (.NET/Lean): C.1 + errors only (execution_count advisory)
QuantConnect notebooks: C.1 + errors only (require QC Cloud for execution)

@github-actions

github-actions Bot commented Oct 6, 2026

Copy link
Copy Markdown
Contributor

✅ No prose/output mismatch detected in the notebooks this PR changed.

Scope = notebooks CHANGED in this PR, not the whole corpus. Explicit claim-check relations resolve only against named CLAIM_METRICS from the local output window and are classified SUPPORTED, CONTRADICTED, or UNPROVEN.
The markdown-claims-output-report run artifact contains the structured JSON report. See python scripts/check_markdown_claims_output.py --help for re-running locally.
Detector rationale: c.290 / c.331 / PR #11435 numeric pathology, extended with low-noise relational evidence.

@github-actions

github-actions Bot commented Oct 6, 2026

Copy link
Copy Markdown
Contributor

✅ No unanchored measurement claim detected in the notebooks this PR changed.

Scope = notebooks CHANGED in this PR, not the whole corpus. The stale-claim-report run artifact holds the structured JSON.
Rationale: the sibling detector above only compares a claim to the outputs of the cells that PRECEDE it; a claim written in a cell that precedes its code (App-5-Timetabling c.2/c.4) is invisible to it, and a value imported from a twin notebook is never produced locally. See python scripts/check_stale_claims.py --help.

@github-actions

github-actions Bot commented Oct 6, 2026

Copy link
Copy Markdown
Contributor

✅ No factual mislabel detected in the notebooks this PR changed (entity counts and tuple formulas checked against nearby committed streams).

Scope = notebooks CHANGED in this PR, not the whole corpus. The factual-mislabel-report run artifact holds the structured JSON.
Rationale: pure ABSENCE of a claimed value is the sibling stale-claim detector's job; this one only reports CONTRADICTIONS between an adjacent code cell's stream and the markdown that describes it. See python scripts/check_factual_mislabel.py --help.

@jsboige jsboige mentioned this pull request Oct 6, 2026
Closed
9 tasks done
@clusterManager-Myia clusterManager-Myia mentioned this pull request Oct 6, 2026
Open
@jsboige jsboige mentioned this pull request Oct 6, 2026
Open
4 of 6 tasks
@github-actions github-actions Bot mentioned this pull request Oct 6, 2026
Merged
@github-actions

github-actions Bot commented Oct 6, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Path-collision (organ #13359/#13615)

Cette PR #19537 (feat(notebook,#19494): Percolation-03 critique -- point critique, β/ν + τ' + p_c(L) → 1/2) touche au moins un chemin de fichier aussi modifie par d'autres PRs ouvertes. Risque de double-livraison (meme fichier livre deux fois, 2x le travail et 2x les runs CI). Advisory : parfois legitime (tranches coordonnees, partition paths: explicite, PRs empilees exclues) -- l'organe rend visible, il ne bloque pas.

@github-actions github-actions Bot added the pr-overlap Advisory: another open PR touches the same files (organ #13615) label Oct 6, 2026
@jsboige @claude
chore(percolation,#19537): empty commit pour re-trigger PR gate post …
bfae06a 
…c.1125

Latest-wins-green: tous sub-checks SUCCESS. PR gate FAILURE = artefact c.1120-N2. Declencheur empty commit.

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@github-actions github-actions Bot removed the variation-adjacency-deep-med Adjacence DEEP/MED hors LIGHT : §2 l'autorise si substance distincte (coordinateur) label Oct 7, 2026
@jsboige @claude
fix(percolation,#19537): prose-counts ARTIFACT_NOUNS '0 cellule' -- f…
56098ec 
…ix inherited pile

Suite relecture c.1134 : le gate prose-counts sur le stack Percolation
(#19537 → #19548 → #19550 → #19556) flagge '0 cellule' dans
docs/research/percolation-03-critique-results.md (ligne 39).

Cause : c.1120-N1 ★★ ARTIFACT_NOUNS 'cellule' -- `0 erreur, 0 cellule`
matche le regex `(?<![\w.#])~?\*{0,2}\d{1,6}\*{0,2}\s+(?:lignes?|...|cellules?|...)`
(#9377 doctrine : supprimer la mesure, garder le prédicat).

Fix : `0 erreur, 0 cellule` → `0 erreur, aucune cellule`. Le prédicat
qualitatif 'aucune cellule' transmet la même exigence sans compteur.

Effet attendu :
- #19537 (cette PR) : prose-counts vert, lève le rouge hérité
- #19548 (c1105) : rebase sur c1104 → prose-counts vert
- #19550 (c1106) : rebase sur c1105 → prose-counts vert
- #19556 (c1107) : rebase sur c1106 → prose-counts vert

Lane myia-po-2023:CoursIA-2, c.1134.

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
jsboige added a commit that referenced this pull request Oct 7, 2026
@jsboige
feat(notebook,#19494): Percolation-03b critique -- extensions L=128, …
d27b324 
…L=256 (convergence asymptotique)

Suite directe de PR #19537 (c.1104 execution L in {8, 16, 32, 64}).
Ce carnet etend le sweep a L in {128, 256} pour verifier la convergence
asymptotique des exposants critiques :

- beta/nu = 0.1065 (cible 5/36 = 0.1389, R^2 = 0.964, std_err = 0.010)
- tau_prime moyen (6 L) = 2.188 (cible 187/91 = 2.055, ecart ~7%)
- p_c(L) -> 1/2 monotone (5 L, ecart 0.0016 a L=256) -- PASS conforme

Verdict honnete : la convergence est en cours, asymptote non atteinte a
L=256. L'extension n'a pas suffi ; L=512 serait necessaire pour
approcher beta/nu a 1%% de la cible.

320 nouvelles simulations, ~48 s wall-clock, 32 seeds.
jsboige added a commit that referenced this pull request Oct 7, 2026
@jsboige
feat(notebook,#19494): Percolation-03c critique -- extension L=512 (a…
55802f3 
…symptote non atteinte)

Suite directe de PR #19537 (c.1104 L=8..64) et PR #19548 (c.1105 L=128,256).
Ce carnet pousse le sweep a L=512 pour verifier la convergence
asymptotique des exposants critiques :

- beta/nu = 0.1062 (cible 5/36 = 0.1389, R^2=0.977, std_err=0.007)
  -- PAS d'amelioration vs c.1105 (0.1065 -> 0.1062 = -0.0003)
  -- R^2 augmente (0.964 -> 0.977) mais pente reste ~25%% sous cible
- tau_prime L=512 = 2.913 (cas pathologique, bulk domine par petites comps)
- p_c(L) -> 1/2 monotone, ecart 0.0013 a L=512 (PASS conforme)

Verdict honnete : la convergence de beta/nu est STRUCTURELLEMENT BLOQUEE
aux corrections d'echelle finie. Le taux est de +0.002 par doublement
de L. Pour approcher 1%% de la cible, il faudrait L=1024 ou une analyse
theorique des corrections d'echelle (scaling corrections).

160 nouvelles simulations, ~186 s wall-clock, 32 seeds.
@jsboige @claude
fix(percolation,#19537): STALE_LINK c.1131-N1 -- _quarto.yml entry + …
06b7578 
….html link pour Percolation-03

Solution canonique c.1131-N1 ★★★ : ajouter Percolation-03-Critique-Python.ipynb
a la liste _quarto.yml render + convertir lien README .ipynb -> .html.

Fix le PR gate FAILURE sur Audit README -> .ipynb links (defaut #13025)
sur la PR #19537 (bottom of stack c1104) qui etait en stack avec c1105-c1107
mais en isolation : sans _quarto.yml entry, le .html ne peut pas etre
render target Quarto = la conversion .ipynb -> .html seule ne tient pas
(cf check-docs-links.py:344 = .html valide ssi sibling .ipynb existe ET
notebook dans _quarto.yml project.render).

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@github-actions

github-actions Bot commented Oct 7, 2026

Copy link
Copy Markdown
Contributor

G-VAR-2/3 GENRE signals (advisory, non bloquant, #10020).
La lane `myia-po-2023:CoursIA-2` voit ces signaux actifs sur les mergees du jour (UTC 2026-10-07) :

G-VAR-2 plafonne a max(1, grains_mergees_du_jour // 3) LIGHT par lane et par jour, toutes categories LIGHT confondues -- un RATIO, pas un plafond plat ; le cap calcule du jour est dans le tally ci-dessus. G-VAR-3 interdit deux genres LIGHT consecutifs. Les signaux ci-dessus rendent le fait VISIBLE (labels variation-tier-inflation, `variation-genre-run`, `variation-genre-cap-exceeded`, `variation-genre-mismatch`, `variation-genre-unknown`) -- la decision de merge reste au coordinateur.

@github-actions github-actions Bot added the variation-adjacency-deep-med Adjacence DEEP/MED hors LIGHT : §2 l'autorise si substance distincte (coordinateur) label Oct 7, 2026
jsboige added a commit that referenced this pull request Oct 7, 2026
@jsboige @claude
chore(percolation,#19537): empty commit pour re-trigger PR gate post …
caad7fb 
…c.1125

Latest-wins-green: tous sub-checks SUCCESS. PR gate FAILURE = artefact c.1120-N2. Declencheur empty commit.

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
jsboige added a commit that referenced this pull request Oct 7, 2026
@jsboige
feat(notebook,#19494): Percolation-03b critique -- extensions L=128, …
4046bcf 
…L=256 (convergence asymptotique)

Suite directe de PR #19537 (c.1104 execution L in {8, 16, 32, 64}).
Ce carnet etend le sweep a L in {128, 256} pour verifier la convergence
asymptotique des exposants critiques :

- beta/nu = 0.1065 (cible 5/36 = 0.1389, R^2 = 0.964, std_err = 0.010)
- tau_prime moyen (6 L) = 2.188 (cible 187/91 = 2.055, ecart ~7%)
- p_c(L) -> 1/2 monotone (5 L, ecart 0.0016 a L=256) -- PASS conforme

Verdict honnete : la convergence est en cours, asymptote non atteinte a
L=256. L'extension n'a pas suffi ; L=512 serait necessaire pour
approcher beta/nu a 1%% de la cible.

320 nouvelles simulations, ~48 s wall-clock, 32 seeds.
jsboige and others added 3 commits October 7, 2026 12:58
@jsboige @claude
fix(percolation,#19537): re-exec papermill 7/7 cells OK 2026-10-07
69bd8a3 
Re-execution Percolation-03-Critique-Python via MCP jupyter-papermill :
- 7/7 cells succeeded, 0 errors, 0 fail
- C.2 compliance (commit AVEC outputs)
- Outputs: β/ν = 0.102 FAIL, τ' = 2.20 FAIL (convention honnête),
  p_c(L) → 1/2 PASS conforme
- Wall-clock ~4 s mono-thread, kernel python3

Garde c.1134-N1 DWELL : la PR gate peut re-echouer en FAILURE avec
tous checks SUCCESS (120 min plancher). Re-exec bump commit_author
et re-declenche CI. Patience 10-15 min post-push pour re-aggregation.

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@jsboige @claude
fix(percolation,#19537): re-order cells 15/16 -- section 8 avant 9 (c…
bcb1ca9 
…ell-order SECTION_ORDER)

Re-execution a inverse les cellules markdown '## 9. Limitations' et '## 8. Synthèse' -- le gate cell-order rejette la numerotation arriere. Swap cellules 15/16 pour '## 8' puis '## 9'.

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@jsboige @claude
fix(percolation,#19537): docs-index-guard -- ajouter memo 03-critique…
e2c7dff 
… a docs/README.md

Le memo `docs/research/percolation-03-critique-results.md` (ce PR) etait absent de la section 'Recherche (docs/research/)' du docs/README.md (c.1120-N3 ★). 1 doc inatteignable -> fast-lane FAILURE -> Always-on guards FAILURE -> PR gate FAILURE.

Fix: ajouter 1 ligne dans la table pour le memo 03-critique, format descriptif qualitatif (pas de ARTIFACT_NOUNS, le predicat 'sweep / instrument / verdict' ne matche pas le regex prose-counts).

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@github-actions github-actions Bot removed the variation-adjacency-deep-med Adjacence DEEP/MED hors LIGHT : §2 l'autorise si substance distincte (coordinateur) label Oct 7, 2026
This was referenced Oct 8, 2026
Merged
Closed
Closed
Merged
Open
Open
Open
Merged
Open
Open
Merged
@github-actions github-actions Bot added the large-pr-no-review PR > seuil sans review (ni bot ni humaine) -- retire quand une review arrive (#11232) label Oct 8, 2026
@github-actions

github-actions Bot commented Oct 8, 2026

Copy link
Copy Markdown
Contributor

Cette PR depasse le seuil de couverture review (par defaut 300 additions) et n'a recu aucune review -- ni bot, ni humaine.

Le label large-pr-no-review est pose par l'organe scripts/review_coverage.py porte par l'issue #11232. Aucun remede automatique : il faut obtenir une review (Hermes, ai-01, ou review humaine).

Le label est retire au balayage suivant (quotidien) des qu'une review arrive -- dans reviews[] ou en commentaire de verdict -- ou que le diff passe sous le seuil. Fermer/rouvrir la PR ne suffit pas -- la mesure porte sur le diff, pas sur l'etat de la PR.

Seuil, historique et exceptions : cf. docs/reference/review-coverage-threshold.md.

This was referenced Oct 8, 2026
Open
Open
Open
Open
Open
Open
Open
@jsboige
Merge remote-tracking branch 'origin/main' into feature/c1104-percola…
e826563 
…tion-03-exec

# Conflicts:
#	_quarto.yml
#	docs/README.md
@jsboige

jsboige commented Oct 8, 2026

Copy link
Copy Markdown
Owner Author

[ADJOINT PREFLIGHT]
schema: 1
lane: myia-po-2026:CoursIA-3
pr: 19537
head: e826563
complete: true
body: read
comments-reviewed: 12
reviews-reviewed: 0
threads-reviewed: 0
threads-unresolved: 0
surfaces-sha256: 49cc1023cd6333346ec4fd32053a099a73fab6652227c7f6765dcd3ac17d7403
diff-files: 5
diff-additions: 1019
diff-deletions: 0
checks: BLOCKED
b0: clear
scope: pass
domain: pass
verdict: BLOCKED
organ: check_adjoint_prevalidation.py
organ-command: python scripts/check_adjoint_prevalidation.py --derive-verdict 19537
organ-rc: 3
[/ADJOINT PREFLIGHT]

jsboige added a commit that referenced this pull request Oct 8, 2026
@jsboige @claude
fix(percolation,#19494): index docs, ordre sections 03, compteurs prose
8b3547d 
Trois causes du PR gate rouge, reproduites localement puis corrigees :

1. docs-index-guard : les cinq docs de la serie Percolation (03, 03b, 03c,
   04-scoping, 04-results) n'etaient atteignables depuis aucun index.
   Ajout des cinq entrees dans la table Recherche de docs/README.md ;
   230/230 docs vivants atteignables apres correctif.

2. Cell-order gate : la copie du carnet 03 portee par cette pile avait les
   deux dernieres cellules markdown permutees (section 9 Limitations avant
   section 8 Synthese). Ordre restaure a l'identique de la version validee
   sur la branche c1104 (PR #19537), dont le gate cell-order est vert.
   Diff limite aux sources de ces deux cellules : zero execution_count et
   zero outputs touches.

3. prose-counts : cinq compteurs quantitatifs en prose retires (0 cellule,
   100 lignes x2, 20 cellules, 1 ligne) au profit du predicat seul,
   conformement a la tenue des donnees par le CI plutot que par la prose.

Co-Authored-By: Claude Haiku 4.5 (1M context) <noreply@anthropic.com>
@jsboige jsboige mentioned this pull request Oct 9, 2026
Open
3 tasks
This was referenced Oct 9, 2026
Open
Open
Merged
Open
Open

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

large-pr-no-review PR > seuil sans review (ni bot ni humaine) -- retire quand une review arrive (#11232) markdown-table-syntax Table syntax defect in changed files (CODE_SPAN_PIPE, NO_SEP, ...). Advisory. See #10097. pr-overlap Advisory: another open PR touches the same files (organ #13615)

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@jsboige